SEMPAI: a Self-Enhancing Multi-Photon Artificial Intelligence for prior-informed assessment of muscle function and pathology
This addresses the need for interpretable and data-efficient AI in biomedical research, particularly for muscle pathology and function analysis, though it is domain-specific and incremental in combining priors with deep learning.
The paper tackles the problem of black-box deep learning in biomedical studies by introducing SEMPAI, a method that integrates hypothesis-driven priors into a data-driven approach for analyzing multiphoton microscopy images of muscle fibers, resulting in outperforming state-of-the-art biomarkers in six of seven predictive tasks.
Deep learning (DL) shows notable success in biomedical studies. However, most DL algorithms work as a black box, exclude biomedical experts, and need extensive data. We introduce the Self-Enhancing Multi-Photon Artificial Intelligence (SEMPAI), that integrates hypothesis-driven priors in a data-driven DL approach for research on multiphoton microscopy (MPM) of muscle fibers. SEMPAI utilizes meta-learning to optimize prior integration, data representation, and neural network architecture simultaneously. This allows hypothesis testing and provides interpretable feedback about the origin of biological information in MPM images. SEMPAI performs joint learning of several tasks to enable prediction for small datasets. The method is applied on an extensive multi-study dataset resulting in the largest joint analysis of pathologies and function for single muscle fibers. SEMPAI outperforms state-of-the-art biomarkers in six of seven predictive tasks, including those with scarce data. SEMPAI's DL models with integrated priors are superior to those without priors and to prior-only machine learning approaches.